Abstract Purpose To compare seed loss and dosimetric parameters between stranded and loose 125I seeds (LS) for prostate brachytherapy. Methods and materials Sixty-four patients with 1997 American Joint Commission on Cancer (AJCC) clinical stage T1c or T2a prostate carcinoma were prospectively randomized to brachytherapy (144 Gy) with RAPID Strand 125I seeds (RS) vs. LS (Oncura, Plymouth Meeting, PA) The treatment plan for each patient was devised before randomization, and was not modified based on the randomization. Each patient underwent magnetic resonance, computed tomography, and plain film radiographs on the day of the implant (Day 0) and 30 days later (Day 30). Results Overall, 21 of 62 patients (30%) experienced seed loss. Seed loss occurred in 15 of 32 of LS patients (47%) vs. 6 of 30 RS patients (23%; p = 0.053). Mean seed loss was 1.09 in the LS patient vs. 0.43 in RS patients ( p = 0.062). Eight LS patients (25%) lost multiple seeds, compared to 3 stranded patients (10%). Despite the lesser degree of seed loss in patients who received stranded seeds, they had a paradoxical trend toward lower V 100 and D 90 values. Conclusion This prospective randomized trial showed a strong trend toward a decrease in postimplant seed loss with stranded seeds. Improved seed retention may be more advantageous in a setting of less generous periprostatic coverage. The lowered risk seed migration seen with stranded seeds would presumably also decrease the likelihood of lung or cardiac seed embolization.